#microformats 2018-07-18

2018-07-18 UTC
[matpacker], [kevinmarks], [manton], [chrisaldrich], [eddie], [jgmac1106], [miklb], tantek, [xavierroy], klez, Kaja____, KartikPrabhu, jalcine, iwaim, ben_thatmustbeme, gRegorLove, GWG, dougbeal, schmarty, wakest, hober, reidab, Zegnat, pniedzielski[m], sknebel, aaronpk, voxpelli, JonathanNeal, twisted`, Phae, Garbee, bigbluehat, edsu, MeanderingCode, ivc, ChanServ, MylesBraithwaite, globbot, bear, jmac, echarlie, chrisaldrich, [davross], barpthewire, [tantek], joeld, [kim_landwehr], [dougbeal], tantek__, [cleverdevil] and [ryan339] joined the channel
#
[chrisaldrich] Looks like the microformats.org server has tipped over again....
#
[chrisaldrich] aaronpk, did you get to the root of the problem last week?
#
tantek__ so to speak
#
aaronpk No but I did add some monitoring tools so I can try to see what caused it
[jgmac1106] joined the channel
#
KartikPrabhu there is a backup now right?
#
aaronpk sknebel’s mirror
#
KartikPrabhu might be good to link that in the chat topic too
[colinwalker], [chrisaldrich] and snarfed joined the channel
#
sknebel if anyone can edit that :P
#
snarfed hi all! when parsing an mf2 p- property, *internal* whitespace shouldn't be collapsed, right?
#
snarfed looks like a recent mf2py change started doing that when there are inner tags, eg <div class="p-name">foo <span>bar</span></div> parses to name: ['foobar']
#
snarfed (cc KartikPrabhu)
#
snarfed (note the space between foo and <span> in the html)
#
aaronpk eep that sounds bad
#
snarfed yup
#
snarfed ok, i'll file a bug
#
aaronpk I'm surprised there wasn't a test that failed cause of that
#
snarfed testing++
#
Loqi testing has 1 karma in this channel (30 overall)
#
snarfed filed https://github.com/microformats/mf2py/issues/112
#
Loqi [snarfed] #112 inner whitespace in p- values shouldn't be collapsed
#
sknebel I'd guess with the whitespace changes the mf2-testsuite wasn't run (since it'll conflict with that all the time) and the tests outside it don't cover that?
#
sknebel (guess!)
#
aaronpk The library has its own tests tho right?
#
snarfed it has CI, so if not, i have to wonder what that's doing :P
#
sknebel > and the tests outside it don't cover that
#
sknebel checking the history if a test was accidentially "fixed" to match wrong behavior
#
sknebel for the whitespace changes, the tests from https://pin13.net/mf2/whitespace.html got added, but they don't cover that
#
snarfed sknebel++ thanks for investigating!
#
Loqi sknebel has 16 karma in this channel (113 overall)
#
gRegorLove I don't think php-mf2 has an explicit test for this either, though it appears to handle it correctly. I'll make a note to add some tests.
#
sknebel we really should work on getting a shared test suite up again
#
sknebel although I guess that's tricky if whitespace is being interated on, but some basics there should always hold
#
snarfed i'm skeptical of our enthusiasm for shared test suites
#
snarfed they're great for interop, but for day to day development, they're not a good replacement for separate, per library unit tests
#
aaronpk I realize this is something that's just assumed to have been done correctly, but I'm surprised that literally none of the tests happened to have that case in them
#
snarfed testing++
#
sknebel gRegorLove: did you implement new whitespace stuff in php-mf2 following zegnats proposal?
#
sknebel I seem to remember something like that, but not sure
#
gRegorLove Zegnat did: https://github.com/indieweb/php-mf2/pull/168
#
Loqi [Zegnat] #168 New algorithm for plain text values
#
sknebel thx
#
gRegorLove I don't think that changed the parse result in this example though
#
gRegorLove snarfed, php-mf2 uses microformats/tests as a secondary test suite, still has local tests. Agreed, I think there should always be a lib-based test suite.
#
sknebel even more reason to take a look at it
#
gRegorLove And honestly, I forget to run the mf/tests most of the time, heh.
#
gRegorLove CI only uses the php-mf2 tests
#
sknebel never said there shouldn't be one, just wondering if we can't cover basics better if we collect them collectively instead of each lib on their own breaking something and adding a test for it
#
sknebel ahhh, regex
#
gRegorLove "now you have 2 problems" hehe
[cleverdevil] and [eddie] joined the channel
#
sknebel I think Zegnat implemented his algorithm "correctly" 2 times, but his writeup of it doesn't match, so if the writeup was followed trusting it'd do the same as the 2 implementations. gotta check that precisely though, since regex
#
snarfed sknebel++ agreed! happy to donate https://github.com/snarfed/granary/tree/master/granary/test/testdata . 46 different sets there.
#
gRegorLove http://microformats.org/wiki/index.php?title=microformats2-parsing&oldid=66871 also impacts Zegnat's algorithm
#
Loqi [Tantek Çelik] microformats2 parsing specification
#
sknebel true, to keep in mind. Zegnat's on holiday right now, I'll follow up with him to update that in his writeup and/or put it in a repo somewhere
#
KartikPrabhu yeah the whitespace algo only tested for the things aaronpk has on his test
#
KartikPrabhu I am sure if I fix this it will break somewhere else.
#
sknebel yeah, looking at it right now.
#
KartikPrabhu and my algo is almost what Zegnat has
#
sknebel the newline counting stuff I didn't recognize, I guess that's from one of the WHATWG dependencies in zegnats?
#
KartikPrabhu yes
#
KartikPrabhu sknebel: I suspect this is the offending line https://github.com/microformats/mf2py/blob/master/mf2py/dom_helpers.py#L97
#
sknebel KartikPrabhu: yes, already there and checking on consequences of touching it
#
KartikPrabhu 3 criteria 1) don't break existing tests 2) don't use values that are all spacesi.e. <span>\t\n </span> should not count right? 3) fix this issue by snarfed ;)
#
KartikPrabhu also thanks for fixing anf finding bugs. too busy to do mf2py right now
#
sknebel I was too busy the past few months to help, now I got some time, so happy to help
#
KartikPrabhu sknebel++
#
Loqi sknebel has 17 karma in this channel (114 overall)
tantek__, snarfed, [chrisaldrich], chrisaldrich, [ryan339], [cleverdevil], [matpacker], [Chris_Lott] and tantek joined the channel; snarfed left the channel
#
gregorlove edited /in-reply-to (+63) "note u-in-reply-to supersedes rel=in-reply-to" (view diff)
#
gregorlove edited /rel-in-reply-to (+151) "note superseded by u-in-reply-to" (view diff)
#
gRegorLove Perhaps a good note for clarifying the /process for proposed properties: https://chat.indieweb.org/dev/2018-07-18#t1531956105564100
#
Loqi [tantek__] aaronpk, re: X-Forwarded-For and the whole mnot RFC about disliking x-, all the examples that drove that conclusion were x- that hung around "too long", or rather, with too slow a process for actually assessing x- experiments and then either rejectin...
jmac, bradenslen and KartikPrabhu joined the channel